Compressive Sensing-Based Wireless Array and Collaborative Signal Processing Method

Sensor array signal processing has shown great potential in target monitoring applications. However, for wireless sensor network nodes, sending raw signal would result in large transmission delay and consume too much energy.To achieve low power consumption, high accuracy and deployment flexibility in target monitoring under the framework of wireless sensor network (WSN), a compressed sampling-based wireless array is presented. With the help of compressive sensing theory, real-time array signal transmission is made feasible on low-power, low-cost wireless platform. In this framework, signals are randomly sampled at a lower average sampling rate and reconstructed at the fusion center, which is supposed to have powerful processing capability. Furthermore, based on the correlation of signals within an array, collaborative array signal reconstruction algorithm is designed using priori knowledge model to reduce redundancy in signal collection. Simulation results show that compressed sampling-based wireless array can perform effective direction of arrival (DoA) estimation for targets without distinct performance decline, and it even outperforms traditional method obviously when data transmission is heavily restricted. It is show that CS based DoA estimation has similar performance with conventional method while requiring about 15% data in transmission. Experiment is also conducted on prototype system with low cost microphones arrays, thus verifying the feasibility of implementation.